Artificial Intelligence: Where Are We At When It Comes To Asthma?
When reading about how rapid testing for asthma may one day be doable because of artificial intelligence, I wondered what other applications artificial intelligence might have where it comes to asthma and respiratory health.
Artificial intelligence (AI) is rapidly developing, especially when it comes to medicine. Where we currently are crowdsourcing data when doctors are unsure about a patient’s diagnosis, in apps such as Figure One (also known to the Reply All podcast as "The Secret, Gruesome Internet for Doctors"), doctors can get second opinions themselves by sharing anonymized patient cases for real-time feedback from other docs wanting to be helpful. It is possible that soon, AI may be able to take a lot of this over.
Because there’s so much guesswork in asthma due to the variability, I’d anticipate there to be a lot of space for AI in the field. At this point, what’s out there and where are we now?
What is artificial intelligence?
Okay, I’m not trying to go too basic here. However, before I started writing about AI, I figured I should read a little definition just so I am sure I know what I’m talking about.
Medical artificial intelligence is a system of complex algorithms. Artificial intelligence must first be taught (known as “machine learning”), which then essentially combines the expertise of doctors, the accuracy of computers, and presumably a growing body of knowledge, to develop the very algorithms that allow the AI to “diagnose”, or at least identify, abnormalities or anomalies.1
The asthma artificial intelligence landscape
We think of AI as more recent innovations like the self-driving car. However, artificial intelligence applications as they relate to respiratory health and disease have been documented in the literature back to at least 19912, using AI to diagnose occupational lung disease, and as recently as an article from December 20193 (recent as I write this, anyways!), in using AI to monitor immune cell movement as an indicator of asthma as noted above.
In between, some applications of use for AI where it comes to asthma and respiratory health include ResApp, an app used to monitor cough sounds and provide diagnostic information about the cause of the cough. In the earlier part of the decade, some work was done on prediction of persistent asthma among children using AI.4 A few years later, a study was done on predicting asthma exacerbations using AI.5 More broadly, a study combining the use of AI and social media to monitor early disease outbreaks is ongoing, potentially allowing for a better response within the healthcare system.6
The future: Artificial intelligence in medical encounters
There is some discussion of the role AI might play in future medical encounters—what some might refer to as “doctor replacement”. A perhaps more reasonable way to look at this is, of course, to make the work of doctors more efficient. If AI can accurately “rule out” pathology (disease) in the majority of patients, leaving only “positive” or “questionable” results for a human physician to go through, I can certainly see this being of massive benefit to the medical system.
Personally, I’d be interested in encountering some artificial intelligence where it comes to my healthcare, but I am not sure how far I would want to go. Wearable device I can use to inform decision making but not be reliant upon? 100%. Fully replacing my doctor? Questionable. However, as artificial intelligence improves and becomes more accessible in everyday life and consumer products, well, my opinions could rapidly change on that.
How do you feel about artificial intelligence playing a role in your asthma care?
Are you currently taking Breo Ellipta?